Foraging Decisions Under Uncertainty in Bumble Bees

Abstract

Bumble bee colonies rely on decentralized decision-making by individual foragers, who must navigate environments filled with uncertainty. Foraging decisions are shaped not only by the average value of resources but also by variability in rewards and limited prior information. Behavioral ecologists have long studied how animals adapt to such challenges, yet it remains unclear how colony-level context, environmental complexity, and individual traits interact to shape decision-making in social insects. In this dissertation, I use lab-reared Bombus impatiens colonies to investigate how bees make foraging decisions when facing stochastic rewards, novel options, and biomechanical constraints. Across four experiments, my dissertation examines whether foragers respond flexibly to uncertainty, explore strategically to acquire information, solve novel problems in complex environments, and learn to land efficiently on unfamiliar flowers. Specifically: (1) The first experiment investigates if B. impatiens workers alter their preference for risky versus consistent rewards depending on colony life stage or size. Colonies in different phases of development (worker production vs. reproductive production) were used to simulate changing energetic priorities. In a controlled foraging task, bees chose between a flower with a consistent sucrose reward and one with a variable reward of equal mean value. I found no evidence that colony stage predicted foraging preference, and only a weak, non-robust effect of colony size. These results suggest that risk-sensitive foraging in bumble bees may not be tightly coupled to colony-level traits, at least under the experimental conditions tested. (2) The second experiment examines whether bumble bees engage in strategic exploration when information could improve future decisions. Using a two-armed bandit task in both a walking Y-maze and a flight arena, I manipulated bees’ prior exposure to each flower type and the number of choices they could expect to make (decision horizon). While models of optimal foraging predict that animals should increase exploration when information is more valuable, bees did not do so. Instead, they preferred familiar or previously rewarding flowers, even when these were less informative. These findings suggest that bumble bees rely on simple heuristics like reward history or familiarity, rather than explicitly valuing information gain in uncertain environments. (3) The third experiment looks at whether foraging routines, individual responsiveness, and environmental complexity influence bees’ ability to solve novel tasks. Bees were exposed to artificial flowers requiring unfamiliar manipulations and were housed in either simple or complex foraging environments. I quantified how consistently each bee followed the same foraging sequence (routine-ness), their responsiveness to novel flower introductions, and whether they landed on and solved the novel tasks. Bees with more rigid foraging routines were less likely to innovate, while environmental complexity increased landing on novel flowers but did not affect problem-solving speed. Responsiveness showed interaction effects with environmental complexity, but no individual trait consistently predicted innovation across contexts. These findings highlight how environmental context and behavioral consistency interact to influence innovation in foraging bees. (4) The fourth experiment explored how flower orientation and structural features affect the landing performance of inexperienced bees. Naïve B. impatiens foragers were presented with artificial flowers that varied in orientation (horizontal vs. vertical) and the presence of a labellum (a protruding structure often thought to aid landing). Despite assumptions that these traits enhance foraging success, I found that landing success was uniformly low among inexperienced bees and was not improved by flower orientation or labellum presence. Some bees preferentially contacted the labellum, but this did not translate into more successful landings. These results suggest that floral morphology alone may not facilitate landing in the absence of prior experience, and that motor learning likely plays a crucial role in effective foraging behavior.